An effective hybrid fractal-wavelet image coder using quadtree partitioning and pruning

We describe a simple hybrid fractal-wavelet coder that demonstrates extremely competitive compression characteristics in terms of rate-distortion curves with minimal computational cost. An energy-threshold criterion is used to perform the quadtree partitioning. The compression is enhanced in two ways: (1) pruning of negligible subtrees and (2) variable length encoding of domain pools that is implicitly built into the quadtree partitioning code. To the best of our knowledge, the rate-distortion characteristics of this coder are surpassed only by highly adaptive partitioning schemes that employ intensive searching of domain pools, hence requiring much higher computational times.